基于无人机的青藏高原鼠兔潜在栖息地环境初步研究

Habitat environment affects the distribution of plateau pikas:A study based on an unmanned aerial vehicle

高原鼠兔(Ochotona curzoniae)是青藏高原生态系统中的关键种之一,其栖息地环境特征对其空间分布的影响一直是重要的研究内容。由于传统地面观测方法耗时耗力,而卫星遥感分辨率低没法识别,因而对高原鼠兔栖息地的研究多为定性且缺少大范围量化研究。无人机航拍为研究高原鼠兔栖息地提供了新的可能,本研究于2015年6月-8月在青藏高原进行了定点定高航拍,共布设了约300个工作地点,获取了约1 800张航拍照片;每张照片覆盖地面26 m×35 m,每个像元覆盖地面约为1 cm^2;通过自主开发的照片处理软件对鼠洞进行自动识别和人工校正,并结合样地的地表温度、植被、土壤水分等信息进行分析。结果表明,1)高原鼠兔洞口密度在不同草地类型中的分布显著不同(P<0.05),高寒草甸高原鼠兔洞口密度显著高于其它草地类型,荒漠高原鼠兔洞口密度最小;2)高原鼠兔最适宜的栖息地环境:NDVI值为0.5~0.6,土壤含水量为20%~25%,生长季陆地表面温度为28.5~29℃;3)高原鼠兔洞口密度与NDVI和土壤水分呈显著正相关关系(P<0.05),但与陆地表面温度呈显著负相关关系(P<0.05)。以上结果说明,植被和土壤水热是影响高原鼠兔选择栖息地的主要因子,高原鼠兔偏向于选择植被条件较好、土壤含水量较高,生长季陆地表面温度较低的区域。

Plateau pikas （Ochotona curzoniae） were considered a keystone species on the Qinghai-Tibetan plat eau （QTP）. Much attention has been paid to the effect of habitat environment on their spatial distribution However, most of the previous studies mainly focused on qualitative description, because pikas＇ holes were dif ficuh to identify by satellite-based remote sensing and traditional field observation was time consuming. High precision photography and low costs make the use of unmanned aerial vehicles （UAVs） a feasible option for in vestigating the spatial distribution of pikas at the plot or even the regional scale. In the present study, we inves tigated pikas＇ burrow density by using aerial photography with light UAVs from July to August 2015. Alto gether, approximately 300 work points were set and 1800 aerial photos were obtained. Each photograph covered 26 m×35 m on the ground and each pixel covered approximately 1 cm^2. Coupled with normalized difference vegetation index （NDVI）, soil water content （SWC） and land surface temperature （LST） data derived from satellite based remote sensing, we analysed the effect of habitat environment on plateau pikas＇ spatial distribution. The results showed that 1） the burrow density of pikas significantly varied with grassland types （P〈 0.05） ; alpine meadows had the maximum burrow density, whereas deserts had the lowest; 2） the most suitable habitat for pikas was 0.540.6 for NDVI, 20%-25% for SWC, and 28.5-29 ℃ for LST, respectively; and 3） burrow density was positively correlated with NDVI and SWC （P〈0.05）, but significantly negatively correlated with LST （P〈0.05）. Our results suggested that NDVI, SWC, and LST were the dominant factors affect ing the distribution of pikas. Owing to the high resolution of the aerial photographs, UAVs provided a novel and efficient method of improving our understanding of the role of pikas over large regions on the QTP.